Merge branch 'develop' of https://github.com/bcgsc/pori_python into feat/DEVSU-2797-remove-gsc-defaults_dustin

Author: dustinbleile

pori_python/graphkb/genes.py

@@ -370,7 +370,7 @@ def get_pharmacogenomic_info(
 
 
 def get_gene_linked_pharmacogenomic_info(
-    conn: GraphKBConnection, source: str
+    conn: GraphKBConnection, source: str = PREFERRED_GENE_SOURCE_NAME
 ) -> Tuple[List[str], Dict[str, Tuple[str, List[str]]]]:
     """
     Return two lists from GraphKB, one of pharmacogenomic genes and one of associated variants.

pori_python/ipr/ipr.py

@@ -268,11 +268,22 @@ def select_expression_plots(
 
 
 def create_key_alterations(
-    kb_matches: List[Hashabledict], all_variants: Sequence[IprVariant]
+    kb_matches: List[Hashabledict],
+    all_variants: Sequence[IprVariant],
+    included_kb_matches: List[KbVariantMatch],
 ) -> Tuple[List[Dict], Dict]:
     """Create the list of significant variants matched by the KB.
 
     This list of matches is also used to create the variant counts.
+
+    kb_matches: the full list of matched kb objects found for the reported variants
+    all_variants: the full list of all reported variants, matched or unmatched
+    included_kb_matches: the list of kb_variant ids to be allowed in the key alterations table;
+        this is all kb_variants if partially matched statements are allowed, or
+        the subset of kb_variants that are conditions for at least one
+        fully satisfied statement condition set, if partially matched statements
+        are not allowed (ie, kb_variants that are not part of any fully satisfied
+        statement condition set are excluded)
     """
     alterations = []
     type_mapping = {
@@ -283,7 +294,12 @@ def create_key_alterations(
     }
     counts: Dict[str, Set] = {v: set() for v in type_mapping.values()}
     skipped_variant_types = []
+
+    included_kbvariant_ids = list(set([item['kbVariantId'] for item in included_kb_matches]))
+
     for kb_match in kb_matches:
+        if kb_match['kbVariantId'] not in included_kbvariant_ids:
+            continue
         variant_type = kb_match["variantType"]
         variant_key = kb_match["variant"]
         if kb_match["category"] == "unknown":
@@ -624,6 +640,20 @@ def get_kb_matches_sections(
     kb_statement_matched_conditions = get_kb_statement_matched_conditions(
         gkb_matches, allow_partial_matches
     )
+
+    if not allow_partial_matches:
+        # remove kb_matches that are not part of any fully matched condition set
+        unique_kb_variant_ids = list(
+            set(
+                [
+                    item['kbVariantId']
+                    for conditionSet in kb_statement_matched_conditions
+                    for item in conditionSet['matchedConditions']
+                ]
+            )
+        )
+        kb_variants = [item for item in kb_variants if item['kbVariantId'] in unique_kb_variant_ids]
+
     return {
         "kbMatches": kb_variants,
         "kbMatchedStatements": kb_matched_statements,

pori_python/ipr/main.py

@@ -477,9 +477,6 @@ def ipr_report(
         gkb_matches = [Hashabledict(match) for match in custom_kb_match_filter(gkb_matches)]
         logger.info(f"\t custom_kb_match_filter left {len(gkb_matches)} variants")
 
-    # KEY ALTERATIONS
-    key_alterations, variant_counts = create_key_alterations(gkb_matches, all_variants)
-
     # GENE INFORMATION
     logger.info("fetching gene annotations")
     gene_information = get_gene_information(graphkb_conn, sorted(genes_with_variants))
@@ -526,6 +523,11 @@ def ipr_report(
         gkb_matches, allow_partial_matches=allow_partial_matches
     )
 
+    # KEY ALTERATIONS
+    key_alterations, variant_counts = create_key_alterations(
+        gkb_matches, all_variants, kb_matched_sections['kbMatches']
+    )
+
     # OUTPUT CONTENT
     # thread safe deep-copy the original content
     output = json.loads(json.dumps(content))
@@ -571,9 +573,10 @@ def ipr_report(
     upload_error = None
 
     # UPLOAD TO IPR
+
     if ipr_upload:
         if not ipr_conn:
-            raise ValueError("ipr_url required to upload_report")
+            raise ValueError("ipr_url required to upload report")
         ipr_spec = ipr_conn.get_spec()
         output = clean_unsupported_content(output, ipr_spec)
         try:

tests/test_graphkb/test_genes.py

@@ -172,7 +172,7 @@ def test_get_pharmacogenomic_info(conn):
     EXCLUDE_BCGSC_TESTS, reason="excluding BCGSC-specific tests (requires CGL loader))"
 )
 def test_get_gene_linked_pharmacogenomic_info(conn):
-    genes, matches = get_gene_linked_pharmacogenomic_info(conn, PREFERRED_GENE_SOURCE_NAME)
+    genes, matches = get_gene_linked_pharmacogenomic_info(conn)
     for gene in PHARMACOGENOMIC_INITIAL_GENES:
         assert gene in genes, f"{gene} not found in get_pharmacogenomic_info"
         for rid, variant_info in matches.items():
@@ -198,7 +198,7 @@ def test_get_cancer_predisposition_info(conn):
     EXCLUDE_BCGSC_TESTS, reason="excluding BCGSC-specific tests (requires CGL loader))"
 )
 def test_get_gene_linked_cancer_predisposition_info(conn):
-    genes, matches = get_gene_linked_cancer_predisposition_info(conn, PREFERRED_GENE_SOURCE_NAME)
+    genes, matches = get_gene_linked_cancer_predisposition_info(conn)
     for gene in CANCER_PREDISP_INITIAL_GENES:
         assert gene in genes, f"{gene} not found in get_cancer_predisposition_info"
 

tests/test_ipr/test_data/expression.short.tab

@@ -25,3 +25,4 @@ KLHL25	27.39	outlier_high	increased expression	100.0	6.97	46.01	100.0	210.66		8.
 ZNRF3-IT1	0.2	no_category								6.75	0.12	34	-1.816	2.225	-2.505	1.744	26	0.623	-1.272	0.2	0.3	0.4	0.5
 PTP4A3	1.33	high_percentile	increased expression	100.0	9.74	99.62	100.0	32.2		6.09	-2.217	87	1.863	2.013	0.286	-0.767	66	-0.091	-1.824	0.2	0.3	0.4	0.5
 ERBB2	0.05	no_category		67.0	0.62	51.71	100.0	1.01		4.8	-0.551	61	-2.009	-2.216	0.147	-2.385	86	2.299	-0.953	0.2	0.3	0.4	0.5
+DPYD	1.12	increased rna expression	increased expression	97.0	4.75	73.38	100.0	50.19		0.99	-0.193	92	2.786	0.537	2.113	1.888	31	0.915	0.861	0.2	0.3	0.4	0.5

tests/test_ipr/test_ipr.py

@@ -10,6 +10,8 @@
     get_kb_matched_statements,
     get_kb_statement_matched_conditions,
     get_kb_variants,
+    get_kb_matches_sections,
+    create_key_alterations,
 )
 from pori_python.types import Statement
 
@@ -483,9 +485,9 @@ def test_germline_kb_matches(self):
         "kbContextId": "#135:8764",
         "kbRelevanceId": "#147:32",
         "kbStatementId": "#155:13511",
-        "requiredKbMatches": ["#159:5426", "#161:938"],
+        "requiredKbMatches": ["#159:54261", "#161:9381"],
         "kbVariant": "BRCA1 mutation",
-        "kbVariantId": "#161:938",
+        "kbVariantId": "#161:9381",
         "matchedCancer": False,
         "reference": "MOAlmanac FDA-56",
         "relevance": "therapy",
@@ -494,6 +496,13 @@ def test_germline_kb_matches(self):
     },
 ]
 
+ALL_VARIANTS = [
+    {"variant": "var1", "key": '1', "variantType": 'mut'},
+    {"variant": "var2", "key": '2', "variantType": 'mut'},
+    {"variant": "var3", "key": '3', "variantType": 'mut'},
+    {"variant": "var4", "key": '4', "variantType": 'mut'},
+]
+
 BASIC_GKB_MATCH = {
     "approvedTherapy": False,
     "category": "test",
@@ -696,9 +705,12 @@ def test_partial_matches_omitted(self):
         for item in input_fields:  # we don't care about these for this test
             item["variantType"] = "test"
             item["kbVariant"] = "test"
+
         gkb_matches = create_gkb_matches(input_fields)
-        stmts = get_kb_matched_statements(gkb_matches)
-        kbcs = get_kb_statement_matched_conditions(gkb_matches)
+        sections = get_kb_matches_sections(gkb_matches, allow_partial_matches=False)
+
+        stmts = sections['kbMatchedStatements']
+        kbcs = sections['kbStatementMatchedConditions']
         assert len(stmts) == 2
         assert len(kbcs) == 1  # X only
         assert kbcs[0]["kbStatementId"] == "X"
@@ -748,6 +760,53 @@ def test_partial_matches_omitted_even_when_var_used_elsewhere(self):
         assert len(kbcs) == 2  # X and Z but not Y
         assert "Y" not in [item["kbStatementId"] for item in kbcs]
 
+    def test_kbvariants_removed_from_set_when_not_part_of_full_conditionset_match(self):
+        """When there is a variant that fulfills one part of a statement's condition set,
+        but isn't part of any fully satisfied condition set,
+        the kbvariant record should be removed from the kbvariants list
+        """
+        input_fields = [
+            {
+                "variant": "A",
+                "kbVariantId": "test1",
+                "kbStatementId": "X",
+                "requiredKbMatches": ["test1", "test2", "test3"],
+            },
+            {
+                "variant": "B",
+                "kbVariantId": "test2",
+                "kbStatementId": "X",
+                "requiredKbMatches": ["test1", "test2", "test3"],
+            },
+            {
+                "variant": "A",
+                "kbVariantId": "test1",
+                "kbStatementId": "Y",
+                "requiredKbMatches": ["test4", "test1"],
+            },
+            {
+                "variant": "D",
+                "kbVariantId": "test4",
+                "kbStatementId": "Y",
+                "requiredKbMatches": ["test4", "test1"],
+            },
+        ]
+        for item in input_fields:  # we don't care about these for this test
+            item["variantType"] = "test"
+            item["kbVariant"] = "test"
+        gkb_matches = create_gkb_matches(input_fields)
+        sections1 = get_kb_matches_sections(gkb_matches, allow_partial_matches=False)
+        kbcs1 = sections1['kbStatementMatchedConditions']
+        kbvars1 = sections1['kbMatches']
+        assert len(kbcs1) == 1  # only fully matched condition sets included
+        assert len(kbvars1) == 2  # therefore, kbvars associated with stmt X are pruned
+
+        sections2 = get_kb_matches_sections(gkb_matches, allow_partial_matches=True)
+        kbcs2 = sections2['kbStatementMatchedConditions']
+        kbvars2 = sections2['kbMatches']
+        assert len(kbcs2) == 2  # all condition sets included
+        assert len(kbvars2) == 3  # therefore, no pruning
+
     def test_partial_matches_included(self):
         """check statements that are only partially supported
         are included when allow_partial_matches=True"""
@@ -779,3 +838,21 @@ def test_partial_matches_included(self):
         kbcs = get_kb_statement_matched_conditions(gkb_matches, allow_partial_matches=True)
         assert len(stmts) == 2  # X and Y
         assert len(kbcs) == 2
+
+    def test_create_key_alterations_includes_only_pruned_kbmatches(self):
+        gkb_matches = create_gkb_matches(GKB_MATCHES)
+
+        sections1 = get_kb_matches_sections(gkb_matches, allow_partial_matches=False)
+        key_alts1, counts1 = create_key_alterations(
+            gkb_matches, ALL_VARIANTS, sections1['kbMatches']
+        )
+
+        sections2 = get_kb_matches_sections(gkb_matches, allow_partial_matches=True)
+        key_alts2, counts2 = create_key_alterations(
+            gkb_matches, ALL_VARIANTS, sections2['kbMatches']
+        )
+
+        # check partial-match-only variants are not included in key alterations when
+        # partial matches is false
+        assert len(key_alts1) == 3
+        assert len(key_alts2) == 4